[GRASS5] problems with [rsv].proj

[Markus Neteler]

On Fri, May 24, 2002 at 07:09:29PM -0400, rgrmill wrote:
> 
> Folks,
> 
> I committed changes to datum_shift.c to test for null values of the datum
> names.  I also committed changes to correct the "dubious behavior" that
> Eric pointed out in r.proj and did the same for v.proj and s.proj.
> 
> I compiled the code and ran the updated r.proj without problems.  I have
> not run v.proj or s.proj.  Further, all my test locations have their datum
> and ellipse specified, so they don't actually test the new code.  In
> brief, more testing is advisable.

Roger,

I have tried r.proj again. It still crashes for a standard GRASS lat/long
location (when generating on GRASS startup).

Contents are:
cat grassdata/europa/PERMANENT/PROJ_INFO
name: Latitude-Longitude
proj: ll
ellps: wgs84

When adding a map datum with g.setproj:
cat grassdata/europa/PERMANENT/PROJ_INFO
name: Latitude-Longitude
datum: wgs84
dx: 0.000000
dy: 0.000000
dz: 0.000000
proj: ll
ellps: wgs84

r.proj no longer crashes. For backward compatibility I suggest to also
support the case when the map datum is absent.

Thanks,

 Markus